Wavelet analysis applied to SAR images to detect atmospheric structures

The aim of this work is to evaluate the possibilities offered by the wavelet analysis in the study of the spatial structure of the radar backscatter, which is linked to the spatial structure of the marine atmospheric boundary layer. Continuous wavelet analysis has been applied to a SAR image of the sea surface, previously analysed by the more classical Conditional Sampling technique. The main problem coped with has been to select, among the large number of wavelet maps produced by the analysis (256), those containing the more energetic backscatter structures. Since the phenomenon imaged by SAR was the convective turbulence, a multiscale and quasi-periodic or intermittent process, a selection of wavelet maps according to their mean energy was unsuccessful. On the contrary, a method based on the calculation of the standard deviation of the less probable highest wavelet amplitudes led to a right selection of the wavelet maps and to a convincing reconstruction of the map of backscatter structures, which resulted similar to that provided through the Conditional Sampling.